In this paper, we examine the role of trade shocks in promoting the diffusion of elementary education in subnational units in Brazil during a period in which they had relative financial autonomy to collect export taxes and spend on public goods. The argument is that trade shocks affect asymetrically the tax revenues of state governments and, thus, their expenditures on elementary education per capita according to what crop mix they had. We then show that states with more egalitarian and democratic institutions use positive trade shocks to invest in education, while the opposite takes place in states with less democratic institutions (e.g., in states that had more slaves). We also show using OLS and instrumental variables that positive trade shocks increased expenditures on education per capita and led to higher literacy rates and to more schools per children. The resulting distribution of human capital across states persists until today.

There is a large literature that aims to explain what determines country risk (defined as the difference between the yield of a sovereign's bonds and the risk free rate). In this article, we contribute to the discussion by arguing that an important explanatory factor is the impact that commodities have on the capacity to pay. We use a newly created database with state-level fiscal and risk premium data (between 1891 and 1930) to show that Brazilian states with natural endowments that allowed them to export commodities that were in high demand (e.g. rubber and coffee) ended up having higher revenues per capita and lower cost of capital. We also explain that the variation in revenues per capita was both a product of the variation in natural endowments (i.e. the fact that states cannot produce any commodity they want) and a commodity boom that had asymmetric effects among states. These two effects generated variation in revenues per capita at the state level thanks to the extreme form of fiscal decentralisation that the Brazilian government adopted in the constitution of 1891, which gave states the sole right to tax exports. We also run instrumental variable estimates using indices of export prices for each state. These estimates confirm our findings that states with commodities that had higher price increases had lower risk premia.